changjonathanc

Below is a consolidated reference to the current Model‑Context Protocol (MCP) features that govern how I think, reason, and format replies—excluding the tool catalog itself.

---

1. Instruction‑layer hierarchy

• System → Developer → User: Every response is shaped by this ordered stack. Higher‑layer rules always override lower‑layer rules.
• Guarantees that user‑level preferences are honored unless they conflict with developer or system‑level safety or policy constraints.

2. Multi‑channel response model

changjonathanc

#!/usr/bin/env python3

import argparse
import base64
import hashlib
import json
import os
import secrets
import sys
import time

changjonathanc

import asyncio
import logging
import os
import pty
import pwd
import subprocess
from contextlib import asynccontextmanager

import uvicorn
from fastapi import FastAPI, HTTPException, Request, Response

changjonathanc

import torch

def precompute_freqs_cis(dim: int, end: int, theta: float = 10000.0, dtype=torch.float32, mod=True):
    freqs = 1.0 / (theta ** (torch.arange(0, dim, 2)[: (dim // 2)].float() / dim))
    t = torch.arange(end, device=freqs.device)  # type: ignore
    angles = torch.outer(t, freqs) # type: ignore
    if mod:
        angles = angles % (2 * torch.pi)
    # Now convert angles to the desired lower precision
    angles = angles.to(dtype)

changjonathanc

"""Generates a document causal attention mask based on a document ID tensor"""

from typing import List, Union

import torch
from torch import Tensor
from torch.nn.attention.flex_attention import _mask_mod_signature, or_masks
from attn_gym.masks import causal_mask

changjonathanc

Run huggingface transformers on M1 Macs, on GPU

• Requirement: macOS 12 Monterey

• First, install a conda env with M1 support

wget https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-MacOSX-arm64.sh
chmod +x ~/Downloads/Miniforge3-MacOSX-arm64.sh
sh ~/Downloads/Miniforge3-MacOSX-arm64.sh

changjonathanc

from transformers import AutoTokenizer, AutoModelWithLMHead
  
tokenizer = AutoTokenizer.from_pretrained("gpt2")

# model = AutoModelWithLMHead.from_pretrained("gpt2")

import re

nums = []

changjonathanc

changjonathanc

changjonathanc

from tqdm.auto import tqdm
import time

class LengthedGeneratorWrapper:
    """wraps an infinite generator with length for tqdm"""
    def __init__(self, infinite_generator, len):
        self.generator = infinite_generator
        self.len = len
    def __len__(self):
        return self.len